ARJ: correct byte accounting and truncation errors #723
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
* `getBytesRead()` could drift from the actual archive size after a full read. * Exceptions on truncation errors were inconsistent or missing. * `DataInputStream` (big-endian) forced ad-hoc helpers for ARJ’s little-endian fields. * **Accurate byte accounting:** count all consumed bytes across main/file headers, variable strings, CRCs, extended headers, and file data. `getBytesRead()` now matches the archive length at end-of-stream. * **Consistent truncation handling:** * Truncation in the **main (archive) header**, read during construction, now throws an `ArchiveException` **wrapping** an `EOFException` (cause preserved). * Truncation in **file headers or file data** is propagated as a plain `EOFException` from `getNextEntry()`/`read()`. * **Endianness refactor:** replace `DataInputStream` with `EndianUtils`, removing several bespoke helpers and making intent explicit. * Add assertion that `getBytesRead()` equals the archive size after full consumption. * Parameterized truncation tests at key boundaries (signature, basic/fixed header sizes, end of fixed/basic header, CRC, extended-header length, file data) verifying the exception contract above.
ppkarwasz
commented
Oct 6, 2025
src/main/java/org/apache/commons/compress/archivers/arj/ArjArchiveInputStream.java
Show resolved
Hide resolved
The static import makes it harder to distinguish calls that need to count bytes from those that do not.
garydgregory
requested changes
Oct 11, 2025
src/main/java/org/apache/commons/compress/archivers/arj/ArjArchiveInputStream.java
Outdated
Show resolved
Hide resolved
src/main/java/org/apache/commons/compress/archivers/arj/ArjArchiveInputStream.java
Show resolved
Hide resolved
…length-computation
garydgregory
requested changes
Oct 15, 2025
src/main/java/org/apache/commons/compress/archivers/arj/ArjArchiveInputStream.java
Outdated
Show resolved
Hide resolved
ppkarwasz
added a commit
that referenced
this pull request
Oct 16, 2025
> [!CAUTION] > **Source-incompatible** (callers may need to add `throws IOException` or a catch). > **Binary-compatible** (the `throws` clause isn’t part of the JVM descriptor). ## Motivation Several `ArchiveInputStream` implementations either - must read/validate bytes up front (e.g., magic headers), or - may fail immediately when the underlying stream is unreadable. Today we’re inconsistent: * Formats **without a global signature** (e.g., **CPIO**, **TAR**) historically didn’t read in the constructor, so no `IOException` was declared. * Other formats that **do need early bytes** either wrapped `IOException` in `ArchiveException` (**ARJ**, **DUMP**) or deferred the read to the first `getNextEntry()` (**AR**, **ZIP**). This makes error handling uneven for users and complicates eager validation. ## What this changes * All archive `InputStream` constructors now declare `throws IOException`. * **ARJ** and **DUMP**: stop wrapping `IOException` in `ArchiveException` during construction; propagate the original `IOException`. * **AR**: move reading of the global signature into the constructor (eager validation). No behavioral change is intended beyond surfacing `IOException` at construction time, where appropriate. For the ARJ format this was discussed in #723 (comment). > [!NOTE] > Version `1.29.0` already introduces source-incompatible changes in other methods, by adding checked exceptions.
ppkarwasz
added a commit
that referenced
this pull request
Oct 16, 2025
> [!CAUTION] > **Source-incompatible** (callers may need to add `throws IOException` or a catch). > **Binary-compatible** (the `throws` clause isn’t part of the JVM descriptor). Several `ArchiveInputStream` implementations either - must read/validate bytes up front (e.g., magic headers), or - may fail immediately when the underlying stream is unreadable. Today we’re inconsistent: * Formats **without a global signature** (e.g., **CPIO**, **TAR**) historically didn’t read in the constructor, so no `IOException` was declared. * Other formats that **do need early bytes** either wrapped `IOException` in `ArchiveException` (**ARJ**, **DUMP**) or deferred the read to the first `getNextEntry()` (**AR**, **ZIP**). This makes error handling uneven for users and complicates eager validation. * All archive `InputStream` constructors now declare `throws IOException`. * **ARJ** and **DUMP**: stop wrapping `IOException` in `ArchiveException` during construction; propagate the original `IOException`. * **AR**: move reading of the global signature into the constructor (eager validation). No behavioral change is intended beyond surfacing `IOException` at construction time, where appropriate. For the ARJ format this was discussed in #723 (comment). > [!NOTE] > Version `1.29.0` already introduces source-incompatible changes in other methods, by adding checked exceptions.
garydgregory
pushed a commit
that referenced
this pull request
Oct 16, 2025
> [!CAUTION] > **Source-incompatible** (callers may need to add `throws IOException` or a catch). > **Binary-compatible** (the `throws` clause isn’t part of the JVM descriptor). Several `ArchiveInputStream` implementations either - must read/validate bytes up front (e.g., magic headers), or - may fail immediately when the underlying stream is unreadable. Today we’re inconsistent: * Formats **without a global signature** (e.g., **CPIO**, **TAR**) historically didn’t read in the constructor, so no `IOException` was declared. * Other formats that **do need early bytes** either wrapped `IOException` in `ArchiveException` (**ARJ**, **DUMP**) or deferred the read to the first `getNextEntry()` (**AR**, **ZIP**). This makes error handling uneven for users and complicates eager validation. * All archive `InputStream` constructors now declare `throws IOException`. * **ARJ** and **DUMP**: stop wrapping `IOException` in `ArchiveException` during construction; propagate the original `IOException`. * **AR**: move reading of the global signature into the constructor (eager validation). No behavioral change is intended beyond surfacing `IOException` at construction time, where appropriate. For the ARJ format this was discussed in #723 (comment). > [!NOTE] > Version `1.29.0` already introduces source-incompatible changes in other methods, by adding checked exceptions.
garydgregory
requested changes
Oct 16, 2025
There was a problem hiding this comment.
Hi @ppkarwasz
Merged #731, please resolve conflicts here. TY!
|
I resolved the conflicts and sorted the methods alphabetically. |
|
@ppkarwasz TY, merged! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
In the current implementation:
getBytesRead()could drift from the actual archive size after a full read.DataInputStream(big-endian) forced ad-hoc helpers for ARJ’s little-endian fields.This PR introduces:
Accurate byte accounting: count all consumed bytes across main/file headers, variable strings, CRCs, extended headers, and file data.
getBytesRead()now matches the archive length at end-of-stream.Consistent truncation handling:
ArchiveExceptionwrapping anEOFException(cause preserved).EOFExceptionfromgetNextEntry()/read().Endianness refactor: replace
DataInputStreamwithEndianUtils, removing several bespoke helpers and making intent explicit.Add assertion that
getBytesRead()equals the archive size after full consumption.Parameterized truncation tests at key boundaries (signature, basic/fixed header sizes, end of fixed/basic header, CRC, extended-header length, file data) verifying the exception contract above.